Re: If the only issue in your decision was simplicity, what high availability(automatic failover ) method would you implament?

I am implementing a standby database to be used as a failover in the event that a primary database goes down. Both databases are on Oracle 8.1.7 on Solaris 8 and the application that is accessing them will be BEA weblogic 6. I have been told that Weblogic will use what it calls multipools which are sets of database connection pools which are used in order, in other words if pool 1 is unavailable, pool2 will be used.  Pool 1 can be pointed to one database and pool 2 to another. The database is a very small, very simple one with a relatively small amount of dynamic data (like, 3 logging tables that a population of probably less than a few thousand people will insert small records into a few times a day) records older than 60 days will be cleaned out every night.

My Problem is that I originally thought I would be setting up the entire failover so I figured I could use a standard standby databse and take care of redirecting to the new host in the event of a failure, but since weblogic will be automatically be looking to the new
connection I need to have the data available in both places. all the time I
understand that replication in the true sense is quite complicated. Does anyone know a more simple way?

Incadently, the hardware is dual processor E450 with .5G ram, so performance is very quick for such a simple system. Thus, my real question as originally posted is one of simplicity. I'm wondering the most simple way to have the standby always up to date, or as near as possible.

Again, any thoughts would be appreciated,

Sean

mickrice_at_techie.com (Mick Rice) wrote in message news:<645b6e08.0110030031.67d3c7dc_at_posting.google.com>...
> Sean,
> You have to be kidding ! Give some decent details of the
> situation you're actually in and then people might be able to help.
> What platform (OS)? What size of database? What user population? What
> database version? What hardware?
> What on earth use is a "...nearly up-to-date (within 30 > minutes)
> standby database..." ?
>
> Mick Rice.
>
>
> sean_denney_at_yahoo.com (sean) wrote in message news:<a6568d08.0110021528.7e4de0f1_at_posting.google.com>...
> > What is the easiest way to have a nearly up-to-date (within 30
> > minutes) standby database that could be automatically failed over to
> > by an application server.
> >
> > Any thoughts appreciated
Received on Wed Oct 03 2001 - 13:00:24 CDT